There is ample evidence showing that the stress of caring for a relative with Alzheimer's Disease (AD) adversely affects caregivers. In the longitudinal study funded by the current grant, we found that caregivers had substantially higher rates of syndromal depressive disorders than controls. Caregivers also had poorer immune function on qualitative immunological assays, and more infectious disease episodes (primarily respiratory tract infections) than noncaregivers. Spousal caregivers' immune function has not returned to the level of well-matched control subjects across the four years of the study completed thus far; moreover, bereaved spousal caregivers to not differ significantly from "continuous" caregivers on immune function or depression after bereavement. In fact, our data suggest that spousal caregivers continue to show immunological down-regulation as well as increased risk for syndromal depressive disorders for at least two years after bereavement. While ample evidence has accumulated demonstrating the adverse mental health consequences of caregiving, few studies have addressed how caregivers adapt after bereavement. The high caregiver-related incidence of syndromal depressive disorders both during and after caregiving is particularly troublesome in light of recent evidence that the long-term outcome of depressive episodes may be much less favorable than previously assumed, with many depressed individuals either failing to recover or frequently relapsing. In the depression literature, an increased relapse rate is associated with an older age of onset for depressive episodes, and impaired social relationships are characteristic of both currently depressed individuals and remitted depressives. Since caregivers have small networks and describe their relationships as less supportive than controls, and since immune function appears to be related to both depression and social support, there are significant reasons for concern about depression, immune function, and health both during caregiving and after bereavement.